OBJECTIVES: Talazoparib and olaparib were evaluated in separate Phase III open-label randomized clinical trials (EMBRACA, NCT01945775; OlympiAD, NCT02000622) in patients with a germline BRCA1/2 mutation (gBRCA1/2mut), HER2- locally advanced and/or mBC. We assessed the feasibility of conducting indirect treatment comparisons (ITCs) by comparing PCT arms of both trials within the subgroup of patients in the 1L mBC setting.

METHODS: We qualitatively assessed baseline characteristics of patients receiving PCT in 1L EMBRACA and OlympiAD subpopulations. Unanchored matching-adjusted indirect comparisons (MAICs) using individual patient data from EMBRACA and published summary-level data from OlympiAD were conducted to compare between PCT arms. Patients in EMBRACA were matched to the eligibility criteria of OlympiAD and reweighted to adjust for remaining imbalances in treatment-effect modifiers. Overall survival (OS) was investigated using the Cox proportional hazard model.

RESULTS: The unmatched, unadjusted PCT arm of EMBRACA yielded higher OS versus the PCT arm of OlympiAD (HR 0.39 [95% CI 0.21-0.70]). When patients from EMBRACA who did not meet OlympiAD eligibility criteria were excluded, the matched population yielded higher OS for EMBRACA’s versus OlympiAD’s PCT arm (HR 0.40 [95% CI 0.22-0.73]). After further adjusting for cross-trial differences in published baseline characteristics, OS remained higher in EMBRACA’s versus OlympiAD’s PCT arm (HR 0.46 [95% CI 0.25-0.86]). Adjustment of between-trial differences in post-1L antineoplastic use was not possible because published OlympiAD subgroup data were unavailable.

CONCLUSIONS: Matched and adjusted OS was higher among 1L patients receiving PCT in EMBRACA compared with those receiving PCT in OlympiAD, demonstrating additional cross-trial differences are impacting OS (eg, post-1L antineoplastic use in the PCT arms which was not accounted for due to data limitations). Hence, performing naïve indirect comparisons and/or anchored MAICs for OS between talazoparib and olaparib in the 1L gBRCAmut HER2- mBC setting may be inappropriate, biased, and undermine robust/valid comparisons.

Funding: Pfizer